Article pubs.acs.org/accounts
Developing DNA Nanotechnology Using Single-Molecule Fluorescence Roman Tsukanov,† Toma E. Tomov,† Miran Liber, Yaron Berger, and Eyal Nir* Department of Chemistry and the Ilse Katz Institute for Nanoscale Science and Technology, Ben-Gurion University of the Negev, Beer Sheva, 84105, Israel CONSPECTUS: An important effort in the DNA nanotechnology field is focused on the rational design and manufacture of molecular structures and dynamic devices made of DNA. As is the case for other technologies that deal with manipulation of matter, rational development requires high quality and informative feedback on the building blocks and final products. For DNA nanotechnology such feedback is typically provided by gel electrophoresis, atomic force microscopy (AFM), and transmission electron microscopy (TEM). These analytical tools provide excellent structural information; however, usually they do not provide high-resolution dynamic information. For the development of DNA-made dynamic devices such as machines, motors, robots, and computers this constitutes a major problem. Bulk-fluorescence techniques are capable of providing dynamic information, but because only ensemble averaged information is obtained, the technique may not adequately describe the dynamics in the context of complex DNA devices. The single-molecule fluorescence (SMF) technique offers a unique combination of capabilities that make it an excellent tool for guiding the development of DNA-made devices. The technique has been increasingly used in DNA nanotechnology, especially for the analysis of structure, dynamics, integrity, and operation of DNA-made devices; however, its capabilities are not yet sufficiently familiar to the community. The purpose of this Account is to demonstrate how different SMF tools can be utilized for the development of DNA devices and for structural dynamic investigation of biomolecules in general and DNA molecules in particular. Single-molecule diffusion-based Förster resonance energy transfer and alternating laser excitation (sm-FRET/ALEX) and immobilization-based total internal reflection fluorescence (TIRF) techniques are briefly described and demonstrated. To illustrate the many applications of SMF to DNA nanotechnology, examples of SMF studies of DNA hairpins and Holliday junctions and of the interactions of DNA strands with DNA origami and origami-related devices such as a DNA bipedal motor are provided. These examples demonstrate how SMF can be utilized for measurement of distances and conformational distributions and equilibrium and nonequilibrium kinetics, to monitor structural integrity and operation of DNA devices, and for isolation and investigation of minor subpopulations including malfunctioning and nonreactive devices. Utilization of a flow-cell to achieve measurements of dynamics with increased time resolution and for convenient and efficient operation of DNA devices is discussed briefly. We conclude by summarizing the various benefits provided by SMF for the development of DNA nanotechnology and suggest that the method can significantly assist in the design and manufacture and evaluation of operation of DNA devices.
■
INTRODUCTION Individual molecules are typically too small to be directly observed using visible light, a fact that has significantly hampered studies of biomolecules and efforts to rationally manipulate molecules into designed structures. Progress in these fields of research is, therefore, strongly dependent on the ability of analytical techniques to provide various, but partial, solutions for this major problem. This is also true in DNA nanotechnology: significant effort is invested in the analysis of the building blocks and final products of DNA-based structures and devices, yet the obtained information is limited. The main analytical tools used in the field of DNA nanotechnology are gel electrophoresis, atomic force microscopy (AFM), and transmission electron microscopy (TEM and CryoEM). Gel electrophoresis, the main tool used in the early days of DNA nanotechnology,1 is excellent for determining the size of © 2014 American Chemical Society
a DNA construct and can provide useful information on structural integrity and stability of a DNA construct, and these abilities can be extended further using fluorescence and radioactive labeling.2 AFM3,4 and TEM5−8 provide excellent two-3 and three-dimensional5−8 structural information and are frequently used for analysis of large DNA structures including DNA origami. DNA origami is usually made of a long DNA strand scaffold and many short DNA strands called staples that self-assemble into two3 or three5 dimensional, defined structures. In addition, recent advances in high-speed AFM (HS-AFM) technology enable acquisition of impressive dynamic information Special Issue: Nucleic Acid Nanotechnology Received: January 21, 2014 Published: May 14, 2014 1789
dx.doi.org/10.1021/ar500027d | Acc. Chem. Res. 2014, 47, 1789−1798
Accounts of Chemical Research
Article
Figure 1. Single-molecule fluorescence diffusion-based sm-FRET/ALEX and immobilization-based sm-FRET/TIRF techniques. (A) Optical setup for diffusion-based experiment. (B) Typical time trajectory and photon bursts. (C) E/S histogram. (D) Optical setup for immobilization-based experiment. (E) Donor and acceptor intensities and E time trajectories of individual molecules and typical E histogram and dwell-time histogram calculated from the time trajectories. (F) The position of individual spots can be determined to achieve super-resolution microcopy.
in the preferable aqueous environment.9,10 These analytical tools, however, are not in situ techniques, in that samples are usually not investigated in the aqueous environment or under conditions in which DNA structures are functional, resulting in two main disadvantages. First, the sample may be damaged during the switch to the measurement environment (gel medium or AFM or TEM surfaces). Second, with the exclusion of HS-AFM, it is difficult and sometimes impossible to measure dynamics with these techniques. Fluorescence-based techniques are the methods of choice for in situ study of conformational dynamics of molecules, and bulk fluorescence was the main tool with which dynamics of DNA devices were studied until recently.8,10−15 Because bulk fluorescence provides only averaged information, however, it often fails to capture the real complexity of a device. Valuable information regarding nonreactive and malfunctioning devices and about side products or any other minute population is often lost in the ensemble. These disadvantages are a major problem for the development of dynamic DNA devices. The single-molecule fluorescence (SMF) approach overcomes these limitations. Structural dynamics and interaction information on DNA, RNA, proteins, and DNA−protein complexes16,17 can be obtained with the SMF technique, and it has increasingly become a leading tool in DNA nanotechnology.18−36 The approach, which draws on different methods,37−42 offers a unique combination of features that enables molecular analysis beyond what is possible with more classical methods, thereby expanding the toolkit available for developing DNA-based devices and particularly for the study of their dynamics. The noninvasive character of the fluorescence method enables in situ measurements and allows the system to be maintained in the favored aqueous environment. Single-molecule Förster resonance energy transfer (FRET) provides subnanometer structural information not available through any other method.
When coupled to alternating laser excitation (sm-FRET/ ALEX),43−45 the technique provides stoichiometric information, allowing determination of the presence or absence of building blocks and device integrity. In addition, in cases in which a molecule fluctuates between states, observation of individual molecules may yield kinetic information without a need for nonequilibrium conditions (initiation of kinetic experiment). The single-molecule immobilization-based total internal reflection (TIRF) technique39,41,46 enables continuous observation of individual molecules and, in conjunction with a flow-cell, enables fast replacement of the surrounding solution to facilitate kinetic measurements and convenient introduction of building blocks and removal of waste products of the operation of DNA machines. Finally, because measurements are conducted on individual molecules, the information acquired reflects the distribution of properties rather than the average, enabling detection of minute populations and identification of side products and malfunctioning and nonreactive devices.18,24,25,28 This ability is significantly strengthened using the ALEX technique, which enables isolation and sorting of subpopulations of interest out of the ensemble, increasing resolution.18,19,23−27 This Account is organized as follows: Diffusion-based smFRET/ALEX and immobilization-based TIRF techniques are briefly explained, and the benefits of each technique and which is preferred for what purpose are discussed. sm-FRET/ALEX is then explained in more detail, and various diffusion-based studies are presented. This is followed by a detailed explanation of the TIRF technique and presentation of various immobilizationbased studies. For reasons of clarity, examples of SMF studies of relatively simple systems, such as DNA hairpins26,27 and Holliday junctions,20 are provided first, followed by introduction of more complex DNA origami30,31 and DNA bipedal motor25 systems. Because the aim of this Account is to demonstrate advantages of 1790
dx.doi.org/10.1021/ar500027d | Acc. Chem. Res. 2014, 47, 1789−1798
Accounts of Chemical Research
Article
SMF to the DNA nanotechnology field and to introduce recent advances, we have not attempted to review all studies that employ SMF, especially those reviewed47−52 before. Super-resolution microscopy studies28,29,31,36 are only briefly discussed.30
The continuous observation of individual molecules allowed by the immobilization-based method enables direct observation of conformational changes and of interactions with molecules freely diffusing in the solution, from which transition rates and association and dissociation rates can be obtained. Based on intensity and the number of bleaching steps, immobilization enables better assessment of the number of fluorophores present on an individual molecular unit than does a diffusion-based experiment. Moreover, the ability of TIRF to reject background emission enables affinity measurement in higher concentration than possible in diffusion-based experiments. Furthermore, by using a flow-cell,38,39 the surrounding solution can be replaced while the sample molecule remains in position and is continuously observed. This is a significant advantage because it enables introduction of various components (e.g., building blocks, DNA fuels, see definition below) and removal of leftovers and waste without the need to transfer the sample to a different environment (e.g., using gel electrophoresis or other filters). Finally, superresolution microscopy is possible only for immobilized samples.28−31
■
ADVANTAGES AND DISADVANTAGES OF DIFFUSION-BASED AND IMMOBILIZATION-BASED METHODS Generally, in SMF experiments samples either diffuse freely in the solution or are immobilized on the coverslip surface; examples of both are discussed in the following sections. In the former, picomolar concentrations of fluorescently labeled molecules (or complexes of molecules) diffuse in and out of the confocal spot providing bursts of photons that are usually collected using high time-resolution single-photon avalanche diode (SPAD, ∼200 ps). As a result, the method provides highresolution snapshots of the state of individual molecules. In the immobilization-based technique, individual molecules are continuously observed using an EMCCD camera; these experiments have lower time-resolution than that of the SPAD (∼1−10 ms) but provide time evolution information (essentially a “movie”) of individual molecules. Diffusion-based (Figure 1A) and immobilization-based (Figure 1B) approaches carry advantages and disadvantages, and each is more suitable to answer certain types of questions. Because the diffusion-based method does not require sample immobilization, difficulties associated with the immobilization process and chemistry and due to possible interactions with the coverslip surface are avoided, and therefore, the experiments are usually faster and easier to conduct. The sample molecules are replaced continually by diffusion, and by use of reducing/ oxidizing systems (e.g., Trolox53) and bright and photostable fluorophores (e.g., CY3B, ATTO-532, ATTO-550, ATTO-647N), photobleaching and blinking can be almost completely avoided (Figure 2A), significantly reducing experimental artifacts.
■
DIFFUSION-BASED sm-FRET/ALEX: THE TECHNIQUE With the ALEX technique43,44 (also called PIE45), two or more lasers,22 each directly exciting a corresponding fluorophore, alternate (Figure 1A). Each burst of photons (Figure 1B) produced during the transit of an individual molecule through the confocal spot (the 3D volume in which molecules are observed, roughly 0.3 μm across and 1.5 μm in length) is analyzed in terms of the FRET efficiency and fluorophore stoichiometry (E and S, respectively). The FRET efficiency, defined as E = ADEX/(DDEX + ADEX), where DDEX and ADEX are the number of donor and the acceptor photons, respectively, recorded during the time the donor laser is on, reflects the donor−acceptor distance. The stoichiometry ratio, defined as S = DEX/(DEX + AEX), where DEX and AEX are the sums of the numbers of donor and acceptor photons recorded during the times the donor laser and the acceptor laser are on, respectively, reflects the fluorophore stoichiometry independently of the E value. The results are placed in a two-dimensional E/S histogram that reports on the distributions of donor/acceptor stoichiometry and donor− acceptor distances in the different sample molecules present in the solution (Figure 1C). Other properties of the burst, such as the total number of photons and the duration, are also calculated and provide information on the size of the molecule and the total number of fluorophores attached to the molecule. The method has been used in a number of DNA nanotechnology studies.18−27
■
DIFFUSION-BASED sm-FRET/ALEX: BASIC EXAMPLES
Sorting and Detecting Minute Populations
A two-dimensional E/S-histogram of a duplex DNA made of donor and acceptor labeled strands is presented in Figure 2A. The S values of the doubly labeled duplex are centered on 0.5 indicting the presence of an equal number of donors and acceptors (in fact, one of each fluorophore). Small amounts of donor-only and acceptor-only populations are observed in S = 0.1 and S = 1, respectively. The E-histogram projections are constructed only from events that have S values around 0.5 (Figure 2A, purple rectangle), rejecting events in which one of the fluorophores is absent or not fluorescently active. This example demonstrates that the ALEX technique is capable of identifying and separately investigating different populations including those that account for only a fraction of the total (
Developing DNA Nanotechnology Using Single-Molecule Fluorescence Roman Tsukanov,† Toma E. Tomov,† Miran Liber, Yaron Berger, and Eyal Nir* Department of Chemistry and the Ilse Katz Institute for Nanoscale Science and Technology, Ben-Gurion University of the Negev, Beer Sheva, 84105, Israel CONSPECTUS: An important effort in the DNA nanotechnology field is focused on the rational design and manufacture of molecular structures and dynamic devices made of DNA. As is the case for other technologies that deal with manipulation of matter, rational development requires high quality and informative feedback on the building blocks and final products. For DNA nanotechnology such feedback is typically provided by gel electrophoresis, atomic force microscopy (AFM), and transmission electron microscopy (TEM). These analytical tools provide excellent structural information; however, usually they do not provide high-resolution dynamic information. For the development of DNA-made dynamic devices such as machines, motors, robots, and computers this constitutes a major problem. Bulk-fluorescence techniques are capable of providing dynamic information, but because only ensemble averaged information is obtained, the technique may not adequately describe the dynamics in the context of complex DNA devices. The single-molecule fluorescence (SMF) technique offers a unique combination of capabilities that make it an excellent tool for guiding the development of DNA-made devices. The technique has been increasingly used in DNA nanotechnology, especially for the analysis of structure, dynamics, integrity, and operation of DNA-made devices; however, its capabilities are not yet sufficiently familiar to the community. The purpose of this Account is to demonstrate how different SMF tools can be utilized for the development of DNA devices and for structural dynamic investigation of biomolecules in general and DNA molecules in particular. Single-molecule diffusion-based Förster resonance energy transfer and alternating laser excitation (sm-FRET/ALEX) and immobilization-based total internal reflection fluorescence (TIRF) techniques are briefly described and demonstrated. To illustrate the many applications of SMF to DNA nanotechnology, examples of SMF studies of DNA hairpins and Holliday junctions and of the interactions of DNA strands with DNA origami and origami-related devices such as a DNA bipedal motor are provided. These examples demonstrate how SMF can be utilized for measurement of distances and conformational distributions and equilibrium and nonequilibrium kinetics, to monitor structural integrity and operation of DNA devices, and for isolation and investigation of minor subpopulations including malfunctioning and nonreactive devices. Utilization of a flow-cell to achieve measurements of dynamics with increased time resolution and for convenient and efficient operation of DNA devices is discussed briefly. We conclude by summarizing the various benefits provided by SMF for the development of DNA nanotechnology and suggest that the method can significantly assist in the design and manufacture and evaluation of operation of DNA devices.
■
INTRODUCTION Individual molecules are typically too small to be directly observed using visible light, a fact that has significantly hampered studies of biomolecules and efforts to rationally manipulate molecules into designed structures. Progress in these fields of research is, therefore, strongly dependent on the ability of analytical techniques to provide various, but partial, solutions for this major problem. This is also true in DNA nanotechnology: significant effort is invested in the analysis of the building blocks and final products of DNA-based structures and devices, yet the obtained information is limited. The main analytical tools used in the field of DNA nanotechnology are gel electrophoresis, atomic force microscopy (AFM), and transmission electron microscopy (TEM and CryoEM). Gel electrophoresis, the main tool used in the early days of DNA nanotechnology,1 is excellent for determining the size of © 2014 American Chemical Society
a DNA construct and can provide useful information on structural integrity and stability of a DNA construct, and these abilities can be extended further using fluorescence and radioactive labeling.2 AFM3,4 and TEM5−8 provide excellent two-3 and three-dimensional5−8 structural information and are frequently used for analysis of large DNA structures including DNA origami. DNA origami is usually made of a long DNA strand scaffold and many short DNA strands called staples that self-assemble into two3 or three5 dimensional, defined structures. In addition, recent advances in high-speed AFM (HS-AFM) technology enable acquisition of impressive dynamic information Special Issue: Nucleic Acid Nanotechnology Received: January 21, 2014 Published: May 14, 2014 1789
dx.doi.org/10.1021/ar500027d | Acc. Chem. Res. 2014, 47, 1789−1798
Accounts of Chemical Research
Article
Figure 1. Single-molecule fluorescence diffusion-based sm-FRET/ALEX and immobilization-based sm-FRET/TIRF techniques. (A) Optical setup for diffusion-based experiment. (B) Typical time trajectory and photon bursts. (C) E/S histogram. (D) Optical setup for immobilization-based experiment. (E) Donor and acceptor intensities and E time trajectories of individual molecules and typical E histogram and dwell-time histogram calculated from the time trajectories. (F) The position of individual spots can be determined to achieve super-resolution microcopy.
in the preferable aqueous environment.9,10 These analytical tools, however, are not in situ techniques, in that samples are usually not investigated in the aqueous environment or under conditions in which DNA structures are functional, resulting in two main disadvantages. First, the sample may be damaged during the switch to the measurement environment (gel medium or AFM or TEM surfaces). Second, with the exclusion of HS-AFM, it is difficult and sometimes impossible to measure dynamics with these techniques. Fluorescence-based techniques are the methods of choice for in situ study of conformational dynamics of molecules, and bulk fluorescence was the main tool with which dynamics of DNA devices were studied until recently.8,10−15 Because bulk fluorescence provides only averaged information, however, it often fails to capture the real complexity of a device. Valuable information regarding nonreactive and malfunctioning devices and about side products or any other minute population is often lost in the ensemble. These disadvantages are a major problem for the development of dynamic DNA devices. The single-molecule fluorescence (SMF) approach overcomes these limitations. Structural dynamics and interaction information on DNA, RNA, proteins, and DNA−protein complexes16,17 can be obtained with the SMF technique, and it has increasingly become a leading tool in DNA nanotechnology.18−36 The approach, which draws on different methods,37−42 offers a unique combination of features that enables molecular analysis beyond what is possible with more classical methods, thereby expanding the toolkit available for developing DNA-based devices and particularly for the study of their dynamics. The noninvasive character of the fluorescence method enables in situ measurements and allows the system to be maintained in the favored aqueous environment. Single-molecule Förster resonance energy transfer (FRET) provides subnanometer structural information not available through any other method.
When coupled to alternating laser excitation (sm-FRET/ ALEX),43−45 the technique provides stoichiometric information, allowing determination of the presence or absence of building blocks and device integrity. In addition, in cases in which a molecule fluctuates between states, observation of individual molecules may yield kinetic information without a need for nonequilibrium conditions (initiation of kinetic experiment). The single-molecule immobilization-based total internal reflection (TIRF) technique39,41,46 enables continuous observation of individual molecules and, in conjunction with a flow-cell, enables fast replacement of the surrounding solution to facilitate kinetic measurements and convenient introduction of building blocks and removal of waste products of the operation of DNA machines. Finally, because measurements are conducted on individual molecules, the information acquired reflects the distribution of properties rather than the average, enabling detection of minute populations and identification of side products and malfunctioning and nonreactive devices.18,24,25,28 This ability is significantly strengthened using the ALEX technique, which enables isolation and sorting of subpopulations of interest out of the ensemble, increasing resolution.18,19,23−27 This Account is organized as follows: Diffusion-based smFRET/ALEX and immobilization-based TIRF techniques are briefly explained, and the benefits of each technique and which is preferred for what purpose are discussed. sm-FRET/ALEX is then explained in more detail, and various diffusion-based studies are presented. This is followed by a detailed explanation of the TIRF technique and presentation of various immobilizationbased studies. For reasons of clarity, examples of SMF studies of relatively simple systems, such as DNA hairpins26,27 and Holliday junctions,20 are provided first, followed by introduction of more complex DNA origami30,31 and DNA bipedal motor25 systems. Because the aim of this Account is to demonstrate advantages of 1790
dx.doi.org/10.1021/ar500027d | Acc. Chem. Res. 2014, 47, 1789−1798
Accounts of Chemical Research
Article
SMF to the DNA nanotechnology field and to introduce recent advances, we have not attempted to review all studies that employ SMF, especially those reviewed47−52 before. Super-resolution microscopy studies28,29,31,36 are only briefly discussed.30
The continuous observation of individual molecules allowed by the immobilization-based method enables direct observation of conformational changes and of interactions with molecules freely diffusing in the solution, from which transition rates and association and dissociation rates can be obtained. Based on intensity and the number of bleaching steps, immobilization enables better assessment of the number of fluorophores present on an individual molecular unit than does a diffusion-based experiment. Moreover, the ability of TIRF to reject background emission enables affinity measurement in higher concentration than possible in diffusion-based experiments. Furthermore, by using a flow-cell,38,39 the surrounding solution can be replaced while the sample molecule remains in position and is continuously observed. This is a significant advantage because it enables introduction of various components (e.g., building blocks, DNA fuels, see definition below) and removal of leftovers and waste without the need to transfer the sample to a different environment (e.g., using gel electrophoresis or other filters). Finally, superresolution microscopy is possible only for immobilized samples.28−31
■
ADVANTAGES AND DISADVANTAGES OF DIFFUSION-BASED AND IMMOBILIZATION-BASED METHODS Generally, in SMF experiments samples either diffuse freely in the solution or are immobilized on the coverslip surface; examples of both are discussed in the following sections. In the former, picomolar concentrations of fluorescently labeled molecules (or complexes of molecules) diffuse in and out of the confocal spot providing bursts of photons that are usually collected using high time-resolution single-photon avalanche diode (SPAD, ∼200 ps). As a result, the method provides highresolution snapshots of the state of individual molecules. In the immobilization-based technique, individual molecules are continuously observed using an EMCCD camera; these experiments have lower time-resolution than that of the SPAD (∼1−10 ms) but provide time evolution information (essentially a “movie”) of individual molecules. Diffusion-based (Figure 1A) and immobilization-based (Figure 1B) approaches carry advantages and disadvantages, and each is more suitable to answer certain types of questions. Because the diffusion-based method does not require sample immobilization, difficulties associated with the immobilization process and chemistry and due to possible interactions with the coverslip surface are avoided, and therefore, the experiments are usually faster and easier to conduct. The sample molecules are replaced continually by diffusion, and by use of reducing/ oxidizing systems (e.g., Trolox53) and bright and photostable fluorophores (e.g., CY3B, ATTO-532, ATTO-550, ATTO-647N), photobleaching and blinking can be almost completely avoided (Figure 2A), significantly reducing experimental artifacts.
■
DIFFUSION-BASED sm-FRET/ALEX: THE TECHNIQUE With the ALEX technique43,44 (also called PIE45), two or more lasers,22 each directly exciting a corresponding fluorophore, alternate (Figure 1A). Each burst of photons (Figure 1B) produced during the transit of an individual molecule through the confocal spot (the 3D volume in which molecules are observed, roughly 0.3 μm across and 1.5 μm in length) is analyzed in terms of the FRET efficiency and fluorophore stoichiometry (E and S, respectively). The FRET efficiency, defined as E = ADEX/(DDEX + ADEX), where DDEX and ADEX are the number of donor and the acceptor photons, respectively, recorded during the time the donor laser is on, reflects the donor−acceptor distance. The stoichiometry ratio, defined as S = DEX/(DEX + AEX), where DEX and AEX are the sums of the numbers of donor and acceptor photons recorded during the times the donor laser and the acceptor laser are on, respectively, reflects the fluorophore stoichiometry independently of the E value. The results are placed in a two-dimensional E/S histogram that reports on the distributions of donor/acceptor stoichiometry and donor− acceptor distances in the different sample molecules present in the solution (Figure 1C). Other properties of the burst, such as the total number of photons and the duration, are also calculated and provide information on the size of the molecule and the total number of fluorophores attached to the molecule. The method has been used in a number of DNA nanotechnology studies.18−27
■
DIFFUSION-BASED sm-FRET/ALEX: BASIC EXAMPLES
Sorting and Detecting Minute Populations
A two-dimensional E/S-histogram of a duplex DNA made of donor and acceptor labeled strands is presented in Figure 2A. The S values of the doubly labeled duplex are centered on 0.5 indicting the presence of an equal number of donors and acceptors (in fact, one of each fluorophore). Small amounts of donor-only and acceptor-only populations are observed in S = 0.1 and S = 1, respectively. The E-histogram projections are constructed only from events that have S values around 0.5 (Figure 2A, purple rectangle), rejecting events in which one of the fluorophores is absent or not fluorescently active. This example demonstrates that the ALEX technique is capable of identifying and separately investigating different populations including those that account for only a fraction of the total (
